February 2019

174 tweets

Google is using the dominance of Chrome and Android to make a profit from destroying journalism. Together with Apple's iOS they enable Facebook to do more of the same. Let's not treat this emergency the way we did climate change.
Quoting a tweet by @fmanjoo ↗
↺ 1
"When was the last time you thought of the internet as a weird and wonderful place?" A beautiful if sad essay from @violetblue. How do we turn this tide? engadget.com/amp/2019/01/31…
↺ 8

Replying to @MaxGendler

Sadly it's an ad...

Replying to @gilgul

We have a bunch of open data positions, send them our way. nytimes.wd5.myworkdayjobs.com/Tech

Replying to @trekonomics

For all the people jumping in to say "cheesecake", that's made which "cream cheese" which isn't cheese.

Replying to @trekonomics

Cheesecake is the best cake. Given it's unrelated to cheese I can only assume it was named as a tribute to all that is good and holy.

Replying to @_danbe and @Dan_Benenson, @trekonomics

It isn't matured. French has similar misnomers like fromage blanc. Those can be desserts.

Replying to @_danbe and @Dan_Benenson, @trekonomics

Many years from now we'll look back on this tweet as the moment Twitter jumped the shark. No. Not a pie. It could be a tart but the crust is the wrong kind. And it contains no cheese. What's next? Pizza from Chicago?

Replying to @_danbe and @Dan_Benenson, @trekonomics

Ricotta is a whey cheese. It's complicated.

Replying to @KKjernsmo and @violetblue

I agree that decentralisation of some sort is the way out, I'm just still waiting for it to actually work. I hope you folks figure it out!
My president is @TheAnnaGat. OK!
Quoting a tweet by @RoseLaoutaris ↗
There are known fixes for the "less informed" bit.
Quoting a tweet by @washingtonpost ↗

Thread of 2 tweets

Replying to @macsym

I meant, like, newspapers.

Replying to @PrivacyMatters and @IABEurope

I was half expecting a claim that incest victims too are entitled to a more relevant ad experience, so in a sense I'm relieved.

Replying to @erinlynnyoung and @arainert

It takes some deliberate curation but there is plenty to read that won't, and for good reason.

Replying to @LAM_Barrett

Ragnarok was brilliant. It's hard to top the badassness of @TessaThompson_x, I hope they make Valkyrie movies.

Replying to @LAM_Barrett and @TessaThompson_x

Yes. It's not entirely clear why they make Marvels without her.
Is the Good Life instagrammable?

Replying to @bechgc and @nytdevs, @arainert, @nicksrockwell, @BenGroCo, @dhalioua, @dgelles, @DavidDWalters, @Clayman, @hamman

I spent a while working from home after my first daughter was born, she became accustomed to contributing to regular W3C working group teleconferences, helping lead the Web to its proverbial full potential.

Replying to @jorgenyc_ and @Jorgehastweets, @MaxGendler, @femalegazebot

Dungeons & Dogs would be pretty awesome though.

Replying to @mnot and @MelindaShore

It's not like the other hills, though. If you die on that one, you get a special URN. Oh yeah. 🎤💧

Replying to @JoeAndrieu and @seanmmcdonald, @PrivacyMatters

There are also attempts at interpreting the GDPR as property rights, primarily in the context of the patrimonialisation of personal data. I'm not saying it necessarily holds, but the turf war is here.

Replying to @kateconger

I would ask @CT_Bergstrom !
The manner in which national bodies vote on ISO standards also hurts. Most countries vote "yes" by default when they don't have time to review a standard, and there are lots of standards. Almost nothing gets rejected. I campaigned to halt a bad ISO standard once, it was hopeless.
↺ 2
That is why, as @fantasai says, the @w3c has a royalty-free policy. It doesn't solve everything but it does remove horse-trading from technical standards and that's huge when you know the alternative.
↺ 1

Thread of 3 tweets

Replying to @seanmmcdonald

Subjectivity should, I would think, be context/question-dependent which ML is not good at. Calling it bias reminds us that it's really just glorified statistics and you'd call a bad estimator biased, but subjective?

Replying to @seanmmcdonald

The difference I have in mind is with respect to the Perspective API. It is used to assess the toxicity of text. If you're text implies that you're a black woman it will rate it more toxic than the same text implying you're a white man. Is it racist?

Replying to @seanmmcdonald

The answer depends on context. For many comments, it skews against various groups and that's deleterious. But it is correct that toxic comments will use race and gender to be toxic, so the mention does indeed increase the odds that such a comment is toxic.

Replying to @seanmmcdonald

So it's biased like a tool that always skews one way. For it to be subjective it would have to skew differently depending on context. Does that make sense? It's hard to summarise in tweets.

Replying to @olivierthereaux

Come to NYC 😊

Replying to @fatemehx2 and @bcrypt, @brave

Just speculation, but what device/OS is this? Wondering what form of id matching could have taken place.

Replying to @zeynep

That fine is a warning shot. If the appeal holds they can issue the same fine every day, and then some. And they can go for remedial action as they did with Vectaury. You're right about consent, but it's not the end of that battle.
"Privacy's Blueprint", by @hartzog. An excellent interweaving of design, privacy, law, tech, and policy; I think you'll find it interesting if you're in any of these fields. I hope this makes it into upcoming regulation.
↺ 5

Replying to @hartzog

Thank *you*! With all the ongoing discussions about what to put in privacy regulations I feel it's going to prove quite useful!
The tattoo is egregious; the whole thread about how big tech is corrupting law schools is a must-read.
Quoting a tweet by @cagoldberglaw ↗
↺ 13

Replying to @LAM_Barrett

It doesn't quite communicate decency.

Replying to @null and @Bouletcorp

Il y a Blendle (launch.blendle.com). À ma connaissance ça ne donne pas des résultats fulgurants, mais bon c’est en beta donc bénéfice du doute.

Replying to @null and @Bouletcorp

Pour le micropaiement, techniquement c’est possible et les browsers vont vers ça. Souhaitable est une question plus difficile. On ne peut pas demander aux gens de payer sans rien donner du contenu — il y a trop de clickbait. Mais pour les news l’intro donne souvent l’essentiel.

Replying to @null and @Bouletcorp

Ça peut aussi être difficile de choisir un niveau de prix, pour beaucoup de raison (notamment que tous les articles ne se ressemblent pas, il y a des risques de concurrencer les abonnements, etc.).
@BrobinBerjon Hey, we were talking about you! Haven't heard from you in forever!
In which @karaswisher is exactly right. nytimes.com/2019/02/08/opi…
↺ 2
Lindsey wouldn’t fall for it because she understands privacy and technology. But how many people out there are being stalked as we speak because they had no idea what to do with the notification?

Thread of 2 tweets

Replying to @trekonomics

I thought the deep time limit (after which tectonics completely renewed everything on the surface) was 3.5Bya? If tectonics started only 1Bya shouldn’t we potentially have traces all the way to the beginning of life? Fascinating stuff.

Replying to @null and @koalie, @BrobinBerjon

Lol, OMG I had no idea it was you!

Replying to @null and @koalie, @BrobinBerjon

Sick burn!

Replying to @gsnedders

Hi Sam, welcome! 🤗❤️

Replying to @jess_devonport

Is there a notable downside to just blocking?

Replying to @jess_devonport

I don't know, it could've hurt the feelings of your sauropod followers? I agree blocking is heavy in this case, but in my (admittedly limited) experience curation is what makes Twitter usable. At least mute?
OH: "You eat quiche like a Brooklynite."

Replying to @LinguaCelta

Isn’t sincerity always defiant?
Yes — I’ve been leading a double life: robertobezjon.com (caught by @beaufour)

Replying to @LinguaCelta

I think defiance can be perfectly gentle! It needs resist and disobey, but there is no need to be unkind about it :)

Replying to @LogicalAnalysis

The bus stop is impressive though.
This article has made me wonder all weekend: has anyone sketched out details of re-insecting the world to prevent insect-driven collapse of the entire food web? Maybe some entomophagy research could provide some useful pointers? theguardian.com/environment/20…
Any tool you particularly like for drawing causal diagrams? Accepting a simple textual format as input is a plus.

Replying to @tobie and @GIPHY

*stares*

Replying to @dcreemer and @PlantUML

Thanks! Checking it out.

Replying to @dcreemer and @PlantUML

So, it doesn’t seem to support causal graphs out of the box. Am I missing an option or do you just repurpose some of the UML diagrams?

Replying to @dcreemer and @PlantUML

Thanks, I'll see if that works best for me. Another option was to try to load Dagitty a JS lib and to see if I could massage the output, but it's not really modern JS...
TFW your colleagues just leave their Emmies lying around...

Replying to @null and @mattmay, @femalegazebot

Entirely appropriate what?

Replying to @Iwillleavenow

Wait. But what if I play Devil's Advocate *for the purpose of delivering you a more relevant ad experience*.

Replying to @fatemehx2 and @natashanyt

If it only put the owner at risk... but it brings about a regime in which lying is fraud and silence suspect. Being paid for data is an idea whose time has come to be buried in the woods and not spoken of again.

Thread of 2 tweets

Replying to @Mimoduso and @MaudObels, @borisschapira

Ici on a "Parent / Guardian 1" depuis un bail (prenant en compte *aussi* le fait que certains s'occupent d'enfants sans être parent) mais tu peux les mettre dans l'ordre que tu veux, s'il y a un prénom féminin c'est le premier appel...

Replying to @fatemehx2 and @natashanyt

Maybe the worst part is that I don’t even know how much you could make if you’re already underpriv’d. Data is largely valued in proportion to what you can sell against it — data of people who can’t buy much would only sell so well. It sure would make the rich instantly richer!

Replying to @fatemehx2 and @natashanyt

What worries me most is the closeted gay kid in high school on whatever social media en vogue that asks for orientation. Either they don’t say and that’s a tell, or they lie and that’s fraud.

Replying to @geoffdelc and @AntonioCasilli

J'adore le diagramme!

Replying to @versac and @geoffdelc, @AntonioCasilli

Carrément!

Replying to @willknight and @techreview

Do you know if that diagram could be made available as a wall poster? Cc @geoffdelc @versac
I dropped out of philosophy and that's about all the formal education I got. The philosophy has actually come in useful to work in an industry that has all the accountability of a beheaded chicken. 🐔
Quoting a tweet by @lindydonna ↗
My 9yo is writing a report on Beyoncé but she won't believe Jay-Z is pronounced "Jay-Zed". Please send evidence.

Replying to @null and @amyvdh

Oh yeah.

Replying to @Simon_Lucy

Indeed!

Replying to @null and @avadacatavra

Jay-Zed!

Replying to @trekonomics

Faut-il qu'ils soient vraiment cons pour réussir à me mettre du côté de Finkielkraut...

Replying to @trekonomics

Je doute qu'il n'en apprenne quoi que ce soit...

Replying to @EvanSchultheis and @JamesForbes17, @mcnees

It doesn't kill so much as millimpede it.
Is New Mexico snow always little snowballs falling from the sky or is that special for me?

Replying to @fatemehx2

Thankfully it's giving you a choice in the matter.

Replying to @fatemehx2

I was being ironic 😏 I think it would be quite interesting to poke under the hood to see what it does with that.

Replying to @fatemehx2

The closer it gets to dystopia the harder it is to tell who has just accepted the situation.
I could stick around...

Replying to @brucel

The Breaker Upperers was lit!

Replying to @alexmilowski and @ndw

How far is it from Santa Fe?

Replying to @null and @ndw, @alexmilowski

I am! It's great!
If I showed up at @sfiscience in the morning and started typing away on my laptop, do you think anyone would notice and kick me out? 😏🙄

Replying to @jimtendo69 and @JimmyDonagee, @TheBJPS

And I can't seem to find it anywhere on the site!
I look forward to @dwallacewells's book. In the meantime, here's a sobering taste. It's encouraging to see Celsius used in The Times, though. Change is possible! nytimes.com/2019/02/16/opi…

Replying to @sfiscience

I use historical models of food web disruptions to try to predict the effect of switching from personal data to contextual data in Internet-wide open auction programmatic systems. Might I fake my way to the coffee machine?😁

Replying to @Iwillleavenow

Love both, but I prefer Marmite which one of these days might lose me my Australian citizenship.

Replying to @Iwillleavenow

If they still make it I highly recommend Cadbury's Vegemite chocolate 🍫.

Replying to @erikphoel

Lol. We can't even get articles published with statically meaningful semantics. Or anything accessible, machine-readable...

Replying to @ewinsberg and @seanmcarroll, @RNesling

Inevitable sooner.

Replying to @hukuzatuna and @sfiscience

Of course, of course, you couldn't pull this off without some sensible opsec 🕵️‍♂️
Today in Facebook Sucks: profiting from a public health crisis.
Quoting a tweet by @juliacarriew ↗

Replying to @derrylm and @kzwa

I second Breaker Upperers.

Replying to @cobun and @Iwillleavenow

Actually, I think it was designed to punk people but then became a sensation 😋 adnews.com.au/news/mondelez-…

Replying to @sfiscience

Yay! Thanks, I'll be sure to drop by next time I'm in town.

Replying to @_abialas and @femalegazebot

It's a draft feature.

Replying to @Cennydd

I'd be curious to hear a while from now if you managed to nail it down and it stuck. When I was consulting it seemed that every time I felt I had a clear description of my offering a new client would come along asking for something that didn't fit.
Consent to open programmatic advertising is the roofie of legal bases. The interesting hidden story here is that these documents make it clear that @IABEurope offered it as a solution to publishers in evident bad faith.
Quoting a tweet by @OpenRightsGroup ↗
↺ 11
The most surprising thing in this story is that there are people out there who still believe there's any brand safety in advertising on YouTube.
Quoting a tweet by @nytmedia ↗
↺ 2
Evidence that the White Walkers made it to the Southwest as late as the 19th century.

Replying to @TheAnnaGat

A lot of these are old. Maybe the boring ones survived?

Replying to @TheAnnaGat

Maybe just a form of selection, boring can be reliable!

Replying to @TheAnnaGat

True! There was money in it back then, especially after 1833 when the ad-based model was invented.

Replying to @TheAnnaGat

HAHAHA! Don't worry, I feel pretty removed from the initial naming decision 😏

Replying to @null and @marcosc

DO IT.

Replying to @Iwillleavenow

Ha, same. Sometimes I do wonder if it's such a good idea, though 😏
I've tried, but frankly I don't think I'll ever get British humour. Take this Brexit thing: they've been going on about it for how long now? Still not funny. It's not very credible anyone would do something that stupid to boot. Maybe time to move on to the next joke?
Open question: if we make it simply illegal for a data controller to pass data on to another data controller, does anything valuable actually break?
↺ 3

Replying to @sandhawke

Please detail! I would expect decentralised apps to remain under the controllership of the user.

Replying to @sandhawke

You can pass on the data as processors. Ask Mastodon to publish for you but not make any secondary use of the data, and same for everyone downstream. There is no reason that they must use data for their own purposes.

Replying to @crschmidt and @sandhawke

That's a good case to look at. Give me one good reason for those to be data controllers as opposed to acting entirely under user direction!

Replying to @StewDuff

I assume that data subjects could be able to pass their data to a controller, otherwise there would be no personal data whatsoever!

Replying to @perigrin

I'm not sure I follow?

Replying to @sandhawke and @crschmidt

Here is a simple intro: ec.europa.eu/info/law/law-t…

Replying to @StewDuff

Good point! I feel that legal and medical are two areas that could relatively easily (in the sense that they're already regulated) have some sort of Article 40/regulator-approved boundaries for this sharing. How do you handle Article 14?

Replying to @JA_Boyle

Indeed, but I feel that medical is an area that could relatively easily (in the sense that it's already regulated) have some sort of Article 40/regulator-approved boundaries for this sharing.

Replying to @JibberJim

Care to develop? I'm not sure I follow.

Replying to @perigrin

I don't understand how those deal in personal data?

Replying to @Al_Faiella

Sorry, I'm not sure what you mean here? How does that create a problem?

Replying to @NickNYC10k

You mean like survey providers? Is there any reason those can't be just data processors?

Replying to @perigrin

You could, but that's where things would get illegal!

Replying to @petetiarks

I mean on any basis whatsoever, except as allowed through an Article 40 set of rules approved by regulators. People have listed a few but not that many. Also, portability is user-mediated, it's not really a controller to controller exchange.

Replying to @Al_Faiella

Sorry but I still don't see how that's an objection. Do you have an example?

Replying to @JibberJim

A bunch of countries seem to be doing fine without, and people don't pay their utilities in advance, that's madness.
TFW you're trying to paint a cat entirely from hearsay.

Replying to @briankardell

I'm pretty sure you wouldn't put the whiskers under the nose!

Replying to @Al_Faiella

Aren't those separate processings anyway?

Replying to @crschmidt and @sandhawke

You don't need to know the implementation details in order for them to be a processor. The key item is that they cannot use the data for their own secondary purposes.

Replying to @JibberJim

I just don't put the French system which is about preventing fraud, checking that banks don't carry out unethical lending, and getting people out of excess debt in the same box as the US system which is about controlling people and incentives to spend money you don't have.
The question "what breaks" is about finding which exemptions are needed. I can find a good exemption regime for the French system, US credit bureaus can die gaping for air.

Thread of 2 tweets

Replying to @petetiarks

The problem with better enforcement is that it has the wrong cost structure. The onus is entirely on complaints and regulators. I think the valid cases for controller to controller exchange are rare enough that we could flip it. Can't until it's proven needed and safe.

Replying to @Al_Faiella

Not any data between entities that happen to have been controllers at some point! Of the same data in a given processing.

Replying to @crschmidt and @sandhawke

If they are controllers they can read your email. They could mine it for all sorts of other uses. It's often an unclear line in the middle, but you generally don't want your email controlled by someone else.

Replying to @crschmidt and @sandhawke

You could absolutely *ask* them to give you more relevant ads. Right now you don't.

Replying to @StewDuff

Ah, I'd forgotten about that (not a piece I need much). Makes total sense, thanks!

Replying to @Al_Faiella

Nope, in fact it is common for processors to require subprocessors, eg. for hosting.

Replying to @danbri

Ha! Excellent naming!
My new profile pic? (via @danbri)
Quoting a tweet by @rudytheelder ↗

Replying to @sandhawke and @crschmidt

It's a fine thing to be pedantic programmers but that's not how the law works! The fiduciary line lacks critique, we're only now starting to poke at it for issues. I'm not sure it'll hold up to the light, we'll see.
Word Art 4ever! ❤️ Don't @ me.

Replying to @petetiarks

You might be right. The goal in asking "what breaks" is to take the simplest approach and to see what needs to be added for it to work. Indeed I suspect outlawing it for LI and consent may suffice (though I can see @IABEurope coming up with a contract framework!).

Replying to @crschmidt and @sandhawke

I believe they shouldn't be, but it depends on what they do with your data. If they mine it to profile you and sell that, definitely. I don't think the primary concern w have is with email though.

Replying to @danbri and @shimown

It's *good* Sir Robin.

Replying to @petetiarks and @IABEurope

Pretty much!
Is there a good “How The Web Works” book/article/course that is suitable for smart people willing to work through the complexities but who aren’t interested in actually learning how to code? I keep having to explain bits and pieces of this, it seems like it would be useful.
↺ 4

Replying to @doriantaylor

Those are good for Web architects, but they won’t tell you things like “this is what the HTTP protocol does” or “here is how cookies work”.

Replying to @null and @StephenEvermore

Yes, that has a lot of useful detail and is a great resource, but at the same time I don’t need all the first bits and a fair bit more detail about HTTP and browser stuff!

Replying to @justinph

Yes, that’s a wonderful article and I am definitely sharing it with some of those folks, but it doesn’t solve the heart of my issue :)

Replying to @lhirlimann

Yeah, I was afraid someone would say that.

Replying to @MarechalPhD and @muira_mccammon

I’m looking for something about how the Web works at the operational level (ie what happens when a user directs their browser to a site). I wish people asked me more about Web governance, but they don’t know it exists ;) But this looks like a good read all the same, thanks a lot!

Replying to @Simon_Lucy and @lhirlimann

No, more about what happens at the HTTP level, what is involved in loading, processing, rendering a page, etc.

Replying to @MarechalPhD and @muira_mccammon

Oooh, thanks! That’s not obvious from the summary. I’ll grab it to take a look!

Replying to @Simon_Lucy and @lhirlimann

Really? How so?

Replying to @kashhill and @je5perl

The two suggestions so far have been twitter.com/MarechalPhD/st… which I intend to check out and “you should write it” which… oh dear god.

Replying to @Simon_Lucy and @lhirlimann

But is that the level that would enable a lawyer, a journalist, a data engineer, an executive, etc. to get how it all hangs together?

Replying to @TzviyaSiegman

Not the history as in Weaving, more of a view of all fits together. Architecture but without cryptic bits. Enough to speak intelligently on the topic but not to make a website.

Replying to @null and @prushforth

Thanks, I'll check it out!
“Third-Party Data Is A Bad Habit We Need To Kick” — Kendell Timmers is VP Advertising Data at @nytimes, I think you’ll want to read what she has to say. adexchanger.com/the-sell-sider…

Replying to @paul_irish and @briankardell, @Simon_Lucy, @lhirlimann

Thanks, that’s really helpful!

Replying to @arthurattwell

That’s a good one I hadn’t seen before, thanks!

Replying to @adactio and @briankardell, @TzviyaSiegman, @timberners_lee

I hadn’t thought of that one but it actually works — thanks!

Replying to @runasand

I don’t think that that will work for the audience I have in mind (you might know them :) but it looks like a good resource, thanks!

Replying to @s010n and @linamkhan

I love how you can’t log into @SSRN unless you disable tracking protection… I guess they’re skeptical of their own fiduciary status. Thanks for the link!
"The Uninhabitable Earth: Life After Warming", by @dwallacewells. This may be the most sobering book I've read (and I enjoyed Jude the Obscure); I think you should read it too. It's worse than you think and sooner than you think, but we can also do more than you think.

Replying to @BrendanEich and @pamelafox, @bcrypt, @flamsmark

To be precise, it’s too much for experts or masochists too; though it’s in reach for some masochistic experts or expert masochists.